Schema graph audit
Field-level review of every JSON-LD block on your top templates, plus the entity relationships that connect them into a single resolvable graph.
What it actually is
I pull the rendered HTML from every top template on your site and extract the JSON-LD the way an AI engine would, not the way Google's Rich Results Test does. Then I score each block at the field level: required fields present, recommended fields present, @id stable across pages, sameAs chains resolving, entity references pointing at things that actually exist in your graph.
Marketers usually run your URL through a validator, see green, and call it done. Validators only confirm syntax. They don't tell you whether your Product entity resolves to the same brand entity across 45,000 PDPs, or whether your Organization @id is stable enough for an AI engine to merge your knowledge across pages. That resolution question is the one that decides whether you get cited.
Deliverables
- Per-template schema map (every @type emitted, every @id, every sameAs)
- Field-level scoring spreadsheet across required, recommended, and AI-relevant fields
- Entity resolution diagram showing where the graph fragments
- List of duplicate, orphaned, or unresolvable entities with file paths
- RankLabs export of the full extracted graph for your engineering team
What breaks without it
When the @graph fragments, AI engines stop merging your entities. Your brand looks like a different organization on every template. Your products lose their brand association. Your articles lose their author identity. The downstream effect is silent: rich results disappear from Google, citations disappear from ChatGPT, and you find out 90 days later in a traffic report.
The most common failure mode I find: a CMS that emits Organization JSON-LD on every template with a different @id each time (often the page URL). The graph technically validates. AI engines see five thousand separate organizations with the same name and resolve none of them.
How it fits the Audit
The graph audit is the spine the rest of the audit hangs off. The AI visibility scan and the revenue impact model both reference findings from this artifact, and the 90-day roadmap orders fixes by what the graph audit surfaces. If you move into a Sprint, this map becomes the blueprint for the @graph architecture work.
The full Audit breakdown
Stop pouring budget into a broken foundation.
If your SEO retainer hasn’t compounded, your AI citations have stalled, or your last technical audit ended in a deck nobody read, that’s not a content problem. It’s an engineering problem. The same engineer who diagnoses ships the fix.